It has been suggested that abnormalities of the immune system play an important role in the pathogenesis of chronic inflammatory conditions such as RA. This study revealed Sema3A expression levels significantly correlated with the histological score of perivascular infiltrates of lymphocytes, focal aggregates of lymphocytes and diffuse infiltrates of lymphocytes. Sema3A is involved in immune responses, and its expression levels are altered in several autoimmune diseases [29, 37, 38]. The NRP1 and PlexinA complex mediates Sema3A signaling in the immune system . NRP1 is expressed by regulatory T cells, a subset of T cells  and PlexinA4 regulates T cell proliferation and activation through inhibition of the actin-cytoskeleton rearrangement and T cell receptor polarization . Recently, it was reported that Sema3A promotes regulatory T cells by enhancing IL-10 production . Regulatory T cells play an important role in maintaining immunological self-tolerance by suppressing autoreactive T cells . Immunohistochemical staining in the current study revealed that Sema3A is expressed in the lining layer. Thus, in OA joints, secreted Sema3A from the lining layer may enhance regulatory T cell functions to suppress autoimmune responses in the sublining layer. In RA, the reduction of Sema3A may abrogate the functions of regulatory T cells, thus allowing the infiltration and focal aggregation of autoreactive lymphocytes in the sublining layer. Further investigation of the correlation between Sema3A and IL-10 expression levels may increase our understanding of their roles in RA.
Catalano reported synovial tissues derived from healthy controls, OA and RA patients exhibited no significant differences although the relative Sema3A expression was lowest in RA samples . This discrepancy might be explained by the difference in sampling numbers, which can influence statistical power, and/or by different qPCR methods used. We quantified Sema3A mRNA expression in a larger number of RA patients (n = 30) and OA patients (n = 23), whereas Catalano used relatively small numbers of RA (n = 10), OA (n = 10) patients and healthy controls (n = 5). For qPCR, we used the standard curve method for relative quantification, whereas Catalano used a comparative Ct method.
Sequential immunohistochemical staining revealed that B cells were observed mainly in the lymphoid follicles, consistent with a previous report . Here we observed that B cells in the follicles expressed NRP1 (Figure 4). Vadasz et al. recently suggested that Sema3A can modulate the autoimmune properties of B cells in SLE . Thus, Sema3A may exert similar functions in NRP1-positive B cells from RA synovial tissues. Vadasz et al. also reported Sema3A serum levels were significantly lower in SLE patients (p < 0.0001) and RA patients (p = 0.047) compared with healthy controls . Since Sema3A is a secreted soluble protein and can enter the systemic circulation [16, 42], lower Sema3A serum levels in RA patients may reflect its reduced expression in knee joints and/or other organs. Additional investigations to quantify the serum levels of Sema3A from OA and RA patients may contribute to further elucidation of pathogenesis of the disease.
It is unclear which factors regulate Sema3A expression in synovial tissues. However, Fukamachi et al. reported that the presence of calcium and histamine could modulate Sema3A expression levels in human keratinocytes and fibroblasts . Since histamine is associated with RA pathogenesis [44, 45], it may be responsible for the decreased Sema3A levels observed in RA synovial tissues.
We found that mRNA expression of VEGF-A was not significantly altered in OA and RA synovial tissues. Hashimoto et al. reported staining intensity for VEGF expression did not differ between RA and OA synovial lining  and Lowin et al. reported that VEGF165 expression did not differ in the chronically inflamed tissue of RA patients and OA patients . In contrast, several reports showed altered expression of VEGF in RA. Lee et al. observed significantly higher levels of VEGF protein in RA compared with OA synovial fluid and serum . Kurosawa et al. observed significant correlations of serum VEGF levels with DAS28-CRP scores . The elevation of VEGF protein in RA synovial fluid and serum may explain the higher total number of VEGF-producing cells in the region . Indeed, our study showed no signification correlation between VEGF-A expression levels and DAS28-CRP scores (Figure 2), but there was a marked increase of synovial tissue thickness of the lining layer in RA (Figure 1). Thus, VEGF levels in RA serum may be increased as previously reported . In an earlier study, Ikeda et al. reported VEGF
was expressed in 41% of RA samples (17 patients) but not in OA samples (8 patients) using reverse transcription-PCR . They also found NRP1 was up-regulated in RA synovial tissues. However, we did not observe significant alterations of VEGF-A (Figure 2B) or NRP1 (Figure 2C) between OA and RA specimens. Kim et al. reported that NRP1 expression was similar in OA and RA using immunohistochemical analyses . These discrepancies might be explained by the difference in sampling numbers, disease duration of the populations studied, extent of inflammation, use of anti-rheumatoid drugs, and/or by different methods for analysis of NRP1 expression levels. Additional investigations in a large sample considering several different conditions are required to clarify these discrepancies.
Although Sema3A inhibits endothelial formation and angiogenesis due to competition with VEGF165 , this study did not identify a significant correlation between Sema3A mRNA expression levels and blood vessel density in RA synovial tissues. This suggests that angiogenesis in RA lesions may be regulated by other mediators, such as placenta growth factor 1, IL-2 and hepatocyte growth factor [47, 52].
The imbalance between Sema3A and VEGF may also affect the etiology of RA. Although VEGF-A expression levels did not exhibit a significant correlation with DAS28-CRP scores, the Sema3A/VEGF-A mRNA ratios demonstrated a relationship with the RA clinical score. Several studies have demonstrated that anti-NRP1 peptides suppressed the survival, adhesion and migration of VEGF165-induced synovial cells, which contribute to cartilage destruction in RA [50, 53]. VEGF165 also increased the production of cytokines by human peripheral blood mononuclear cells . Sema3A may inhibit the action of VEGF165 on synovial cells and inflammatory cells in a competitive manner.
Innervation is important in the pathogenesis of arthritis. Primary afferent sensory nerve fibers are proinflammatory, whereas sympathetic nerve fibers are anti-inflammatory . In RA synovial tissues, numbers of sympathetic nerve fibers decreased while sensory nerves increased when compared with OA . The loss of sympathetic nerve fibers in RA may be caused by increased Sema3C and its soluble receptor NRP2 [6, 55]. However, the mechanism of increased sensory nerve fibers in RA synovial tissues is not fully understood. Decreased Sema3A expression in RA may facilitate the increase of sensory nerves over sympathetic nerves in inflamed RA synovium .
Several limitations of our study should be addressed. First, this study population may be influenced by environmental factors as well as genetic factors. Second, patients enrolled in our study were in the advanced stages of RA disease (mean duration 16.3 years). Thus, our data do not reflect the pathogenesis of RA during the early stages. Third, anti-rheumatoid drugs may alter the expression level of Sema3A and other molecules. In this study 12 patients (40%) received methotrexate (MTX) (mean prednisolone daily dose 2.9 mg), 4 patients (13.3%) received TNF-α blockade (3 patients in combination with MTX, 1 patient in combination with sulfasalazine, mean prednisolone daily dose 3.6 mg), 11 patients received Disease Modifying Antirheumatic Drugs (DMARDs) without MTX (mean prednisolone daily dose 1.6 mg), and 3 patients received prednisolone alone (daily dose 6.2 mg). Although there was no statistically significant difference in Sema3A expression, DAS28-CRP, or Rooney’s inflammation score between the MTX–treated group and other groups (data not shown), we cannot rule out the possibility of altered Sema3A expression by anti-rheumatoid drugs, because each group sample size is small in this study.